The possible co-evolution between galaxies and their central supermassive black holes is supported by the similarity in shape between the Star Formation Rate Density (SFRD) and Black Hole Accretion Rate Density (BHARD) out to z$\sim$ 3. This apparent connection between BH growth and star formation is only established globally; while both trends peak at z$\sim$ 2, the amount of stellar and black hole mass assembly occurring within the same galaxies is unknown. Computing these trends for the same galaxies will mitigate the present sample mismatch and can be accomplished with an IR-selected sample; however, the approach relies on a robust understanding of broadband UV-FIR SED fitting to reliably decompose AGN and SF luminosities over a range of AGN strengths. UV-FIR SED fitting is an effective way to disentangle emission between star formation (SF) and active galactic nuclei (AGN) in galaxies with sufficient broadband photometry. The precision of this approach, however, is affected by the sparse mid-IR data presently available in statistically large samples and the variety of AGN SED shapes predicted by radiative transfer torus models. Given these constraints, the use of SED fitting to characterize supermassive black hole accretion is more uncertain in composite AGN/SF galaxies, when the mid-IR SED is \textit{not} overwhelmed by AGN emission. This is significant, as composite galaxies are ubiquitous in nature ($\sim$ 50-70\% of IR-selected samples) and may represent either weak, low-luminosity AGN or a more luminous AGN population that is heavily dust obscured. In this thesis we construct the SFRD and BHARD for the same sample of galaxies, selected at 250$\mu$m in the COSMOS field, testing a variety of AGN SED decomposition methods and AGN templates. We find that employing optically thick AGN torus models results in a BHARD that is higher than current X-ray observations predict but is more similar to the Compton-Thick BHARD predicted by X-ray population synthesis models. This large population of luminous obscured AGN, revealed by SED Decomposition, results in a BHARD trend that drops more rapidly with decreasing redshift than the corresponding SFRD. Our results imply that universal mass assembly via SMBH growth and SF are not directly linked to grow their mass at similar rates across similar epochs.